© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 63411
    Prev
    Next
    INDEX
    Explanations

    expressions indicating time duration or the effort required for a process

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ayıp
    -0.15
    ostel
    -0.15
    azar
    -0.15
    itag
    -0.14
    ichier
    -0.14
    ény
    -0.14
    irez
    -0.14
    olio
    -0.14
    ollo
    -0.13
    onga
    -0.13
    POSITIVE LOGITS
     before
    0.60
    before
    0.51
     Before
    0.45
     antes
    0.45
    Before
    0.41
    -before
    0.39
    _before
    0.38
    	before
    0.38
     BEFORE
    0.36
    .before
    0.35
    Activations Density 0.029%

    No Known Activations